JP2003162607A - Questionnaire item creating apparatus, method and questionnaire performing apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of supporting appropriately a process to convert questionary items against to objected problems from a description form into a selection form. <P>SOLUTION: The system creates the questionnaire items composed of the questionary items in the description form and in the selection form (11) to be stored (12). It creates a questionnaire form based on information in a storage part (12) and storage parts (15) and (18) described later to try a questionnaire (13), be summarizes answering results of the questionnaire (14) to be stored (15). It performs natural language processing and statistical analysis processing to the results of summarizing of questionnaire results corresponding to each questionary item in the description form (16), creates the questionary items equivalent or relating to the questionary items in the selection or description forms by using information regarding to each questionary item (17) and records in an appropriate form contents and forms of created detailed questionnaire items and information regarding to the questionnaire items becoming sources of contents to create the questionnaire items (18). <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a questionnaire item generation device, a questionnaire item generation method, and a questionnaire execution device and method, and more specifically, conversion of a description form of a questionnaire item into a selection form and a peculiar to a respondent. Regarding the generation of forms and repeated answers of only difference items.

[0002]

2. Description of the Related Art Japanese Patent Laid-Open No. 2000-231527 and Japanese Patent Laid-Open No. 2000-235563 are known as questionnaire-taking devices using the Internet. These are devices that carry out and analyze one survey. Japanese Patent Laid-Open No. 11-28828 discloses a method that envisages multiple questionnaires, but the first time is considered as a test trial, and the second and subsequent times are considered to be the main implementation.
Respondents will all answer the same questionnaire.
In such a framework, a plurality of sets of respondent groups must be prepared, one for test and one for this implementation, and if there are few respondents who satisfy the target condition, execution is impossible.

Japanese Patent Laid-Open No. 2001-109735 discloses an invention that gives a respondent a reward for answering a questionnaire, but its purpose is to issue and send electronic money for conducting commercial transactions on a network. Alternatively, in other inventions, many of them relate to points given by questionnaire sites. For example, the reward is non-financial, and instead of electronic money that can be exchanged for various goods and services, prizes and services provided by companies are paid for. In addition, ensuring the reliability of electronic money is an issue, and at present, each company provides its own point service.

The form of question items used in the questionnaire survey can be roughly classified into a selection form and a description form, and a questionnaire form is designed by combining these two types of items. In addition, the question items in the description format are those in which the respondents freely write "text" to answer the question, for example, "Please write the image you hold for product A". It is of the form. In addition, although the question items related to the profile of the respondent's name, age, social status, etc. can be expressed in a descriptive format, these should be treated as different from the question items of the target task.

Generally, the answer result of the questionnaire is in the form of data (hereinafter arithmetic data) to which various arithmetic operations such as arithmetic operations such as numerical values and time can be applied, categorical data and free description sentence. Considering analysis between question items using answer results, first, for arithmetic data and categorical data, analysis methods for existing categorical data such as cross table analysis, correspondence analysis, and quantification analysis are used. It is possible to (It is obvious that arithmetic data can be easily converted into categorical data.)

On the other hand, in order to use the analysis method for the categorical data, the free description sentence data must first be converted into categorical data by some method. However, it is conceivable that the analyst may directly carry out the analysis visually without using the existing statistical analysis method, but it is easily expected that this will be a heavy work for the analyst. In addition, with the recent spread of the Internet, etc., questionnaire surveys using these technologies are becoming more frequent, but questionnaire surveys using the Internet, etc. are compared to conventional manual surveys. Since it is possible to collect a huge amount of response results at low cost, it can be said that it is impossible for the analyst to directly directly analyze the data of the free description text as described above.

[0007] To put it to the extreme, in order to analyze each question item using the results of the questionnaire survey using the Internet, etc., (1) a questionnaire survey is conducted using a questionnaire form consisting of questionnaire items of all selection formats,
Use existing statistical analysis methods. (2) Conduct a questionnaire survey using a questionnaire form consisting of questionnaire items of both selection format and description format, and use some method to categorize the response results of text description format question items. There are two possible ways to use the existing statistical analysis method after converting the data.

When the latter analysis method is selected, a document classification technique (for example, Japanese Unexamined Patent Publication No. Hei 7 (1999) -134) is used as a method for converting free description sentence data in the description format into categorical data.
No. 36897, JP-A-11-29655, etc.) may be used, and by using this technique, free description text data such that similar free description texts by a specific standard belong to the same category. It becomes possible to convert into various categorical data. However, as can be seen from the results of general questionnaire surveys, the answers to question items in the text description format have some missing data due to the fact that they are voluntary entries, and there are qualitatively significant differences between respondents. It is often done. Furthermore, all categories (options) that make up an item (question item) are often treated as categorical data that are semantically contradictory to each other, but this causes a big problem. In other words, in the choice-type question item, since the choices are included in the presented instructions, the respondent considers the answer within the range of the choices, but in the description-type question item, the choices Does not contain a choice,
Respondents think of their answers without restrictions on their choices.

Therefore, the categorical data generated from the results obtained from the text description type question item is considered to be the answer result to the question item presenting incomplete instructions, and therefore the reliability of the data is very high. I have to say that doubts remain. In other words, considering the validity of the analysis results in the process of conducting an analysis between question items using the response results of the questionnaire, the questionnaire survey was conducted using a questionnaire form that does not include the question items in the text description format. We can obtain the hypothesis that it is desirable to collect some categorical data.

Now, let us consider the situation in which a question item of a questionnaire form is generated. When designing a question item for an issue to be investigated, the question item creator should be able to easily structure it if he / she has detailed knowledge of the subject item. It is considered that it is relatively easy to express it as a form. On the other hand, if the creator of the question item only has vague knowledge about the target problem,
It is difficult to structure it, and as a result, it is thought that the question items for the target task must be expressed in text description format. Under these circumstances,
By providing a mechanism that appropriately assists the questionnaire form designer in structuring the answer results to the text description question items, the questionnaire form designer gains well-structured knowledge of the target problem. In such a situation, it is considered that it is easy to express the question item for the target task in the selection format.

Further, structuring of question item answer results in the text description format described above (structuring of knowledge for the target problem)
Is expected to become a heavy task for analysts, but this problem is not structured by a single operation of information (result results), but by dividing the structuring step into multiple stages. I think it can be reduced. In other words, it is a mechanism to repeat the process of creating a question item in multiple related selection formats and descriptive formats by conducting a questionnaire survey using a questionnaire form including text description format question questions, totaling the response results, and analyzing the response results. By repeatedly using it together with the structure support mechanism for the question form in the description format described above, the analyst gradually selects the question form in the text description format from the selection format that is considered to be equivalent or similar to it. It is considered that it can be converted into question items, that is, structuring of the target problem can be realized.

[0012]

Therefore, the present invention appropriately supports structuring of the answer result of a text description type question item, thereby enabling a plurality of description types and selections related to the question item. By providing a mechanism to support the recursive generation of a question item in the form of a questionnaire, the author of the questionnaire form can finally generate a questionnaire form that does not include the question item in the (text) description format. The first purpose is to do so.

In order to achieve the first object, the present invention is a questionnaire item comprising one or more text description type question items, 0 or more numerical description type question items, and 0 or more selection type question items. , A questionnaire item recording unit that records the questionnaire items generated by the questionnaire item generating unit in an appropriate format, a questionnaire item recorded by the questionnaire item recording unit, and a detailed questionnaire item record described below. Part of the questionnaire, and the questionnaire response result recorded in the below-mentioned questionnaire answer result recording section creates a questionnaire form based on the information of the questionnaire response result, and the questionnaire trial section that tries the questionnaire, and the response results in the questionnaire trial section The questionnaire response result totaling section and the questionnaire response result totaling section A description that applies natural language processing and / or statistical analysis processing to the response results of the corresponding questionnaire result aggregation unit for each of the questionnaire response result recording unit that records the results in an appropriate format and the text description format question item By appropriately using the format question item analysis unit and the analysis information of each question item generated by the description format question item analysis unit, a selection format and / or a text description format equivalent to or related to the question item and And / or the detailed questionnaire item generation unit that generates one or more question items in the numerical description format, the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit, and the source that generates the questionnaire item. And a detailed questionnaire item recording section that records information about the questionnaire items that will be Including the question that can be expressed only in the text description format by providing the questionnaire item generation device that can execute the questionnaire item generation unit and the questionnaire item recording unit only once, and can repeat the other processes once or more. I try to support the process of making items selectable.

However, according to the invention for achieving the first object, the questionnaire item is corrected by repeatedly trying the questionnaire based on the generated questionnaire item.
Mainly to support the creation of a new question item that can be expressed only in the text description format, and to convert it into an equivalent / similar selection question item.
No specific reference is made to the framework of the questionnaire survey implementation system.

However, the repeated questionnaire survey is not performed only for converting the question item in the text description format into the question item in the selection format, but the answer result of the question item in the selection format obtained in the process is It is more efficient and economical to consider it as part of the results of a formal survey. From this perspective, how to answer all valid (choice-type) question items necessary for subsequent analysis from individual survey respondents simply and without burdening the respondents. It is an important issue whether it can be extracted. As a simple solution to this problem, existing respondents will be provided with the answer results of the question items that have already been answered, and only new question items will be answered in the next questionnaire response process, and new respondents will be answered. May provide a mechanism for making all valid question items (existing respondents have answered the selected question items + new question items) answered. In other words, by introducing the above mechanism into the questionnaire implementation system,
We believe that it is possible to avoid having both the already-respondents and the new respondents make redundant answers, and as a result, it is possible to realize a questionnaire implementation system with less burden on the respondents.

Therefore, in the present invention, as described above,
By repeating the questionnaire based on the generated questionnaire items, the questionnaire items are modified and newly created, and the question items that can be expressed only in the sequential description format are converted into the equivalent / similar selection format question items. In a questionnaire survey analysis system (apparatus) using a questionnaire item generation device, relationship information between a question item in an arbitrary text description format and a question item that is equivalent or similar to the question item that is generated later, and an arbitrary information for the generated question item. A second object is to provide a questionnaire implementation system that can realize a questionnaire survey that is efficient and has a low burden on the respondents by appropriately managing information regarding the respondents' answers.

A questionnaire survey is conducted using a questionnaire form consisting of questionnaire items of both selection format and description format, and the response results in the description format are converted into categorical data by some method and then the existing statistical analysis method is used. When selecting the analysis method to be used, the answer to the question item in the descriptive format as seen in the general questionnaire survey results is missing data due to voluntary entry, and qualitatively between respondents. There are often noticeable differences. In addition, although all items that make up a category are generally designed to be semantically contradictory to each other,
This is not guaranteed in the category-item structure generated using the above method. Therefore, even if the free description sentence data is converted into categorical data simply by using document classification technology, it is easily expected that the categorical data obtained will have a large qualitative question. It In other words, considering the validity of the analysis results in the process of conducting an analysis between question items using the answer results of the questionnaire, the questionnaire survey is conducted using the questionnaire form only for the questionnaire items of the selection form, and the categorical answer is given. It is desirable to collect results.

Considering again the situation in which a question item is generated, the question item creator, when designing the question item for the problem he / she wants to investigate, has detailed knowledge of the target problem, if any. Since it should be easy to structure, it is considered relatively easy to express the question item for the target task as a selection form. On the other hand, if the creator of the question item only has vague knowledge about the target problem, it is difficult to structure it, and as a result, the question item for the target problem can be expressed in a description format. It is thought that it is unavoidable. However, by appropriately analyzing the answers to the question items in the description format, the question item generator can acquire sufficient knowledge about the target task. In such a situation, the question item It is easy to express in a selection format.

That is, for a problem that can be expressed only in the descriptive format without considering the results of analysis to be collected and analyzed by executing the questionnaire once, collect the answer results in the descriptive format and answer them. By appropriately analyzing the results, you can generate question items in the selection format for the target problem,
By conducting a questionnaire survey again using the generated question items in the selection format, all answer results can be collected as categorical data (including arithmetic data).
By using the existing statistical analysis method for this answer result, it is possible to perform efficient and valid analysis between question items.

As described above, the questionnaire item is corrected and newly created by repeatedly trying the questionnaire based on the generated questionnaire item, and the question item which can be expressed only in the sequential description form is equivalent / similar to the selection item. If you want to support the conversion of the question items into the question items, do not perform the repeated questionnaire survey only to convert the question items in the descriptive format into the question items in the selective format. It is more efficient and economical to consider that the answers to the question items in question are part of the official questionnaire survey results. From this perspective, how to answer all valid (choice-type) question items necessary for subsequent analysis from individual survey respondents simply and without burdening the respondents. It is an important issue whether it can be extracted.

As a simple solution to this problem, existing respondents are provided with the answer results of already answered question items in the next questionnaire answering process, and only new question items are answered and new answer items are given. It may be possible to provide the respondent with a mechanism that allows all valid question items (selection-type question items already answered by existing respondents + new question items) to be answered. In other words, by introducing the above-mentioned mechanism into the questionnaire implementation system, it is possible to avoid having both the existing respondents and the new respondents make redundant answers, and as a result, the questionnaire that is less burdensome for the respondents. We believe that the implementation system can be realized. However, while this is very effective in data analysis, it also has the problem of reducing motivation for respondents. That is, it is necessary to once again answer a question similar to the question that has been answered. If the motivation for the respondents to the questionnaire decreases, the reliability of the response data may decrease, which is a serious problem for the survey.

Therefore, in the present invention, the motivation of the respondents is maintained by paying a reward for the answers of the questionnaire, particularly by paying a reward more than usual for the answers to the second and subsequent questions. The third purpose is to obtain highly reliable data.

[0023]

SUMMARY OF THE INVENTION The present invention is a questionnaire for generating a questionnaire item including one or more text description type question items, zero or more numerical description type question items, and zero or more selection type question items. The item generation unit, the questionnaire item recording unit that records the questionnaire items generated by the questionnaire item generation unit in an appropriate format, the questionnaire item recorded by the questionnaire item recording unit, and the detailed questionnaire item recording unit described later. A questionnaire that generates a questionnaire form based on the questionnaire items and the information on the questionnaire response results recorded in the questionnaire response result recording section described below, and a questionnaire trial section that tries the questionnaire and a questionnaire that aggregates the response results from the questionnaire trial section Appropriately the result aggregated by the response result aggregation unit and the questionnaire response result aggregation unit For each of the questionnaire response result recording section to be recorded in a format and the question description in the text description format, a description format question item analysis section that applies natural language processing and / or statistical analysis processing to the response results of the corresponding questionnaire result aggregation section By appropriately using the analysis information of each question item generated by the above-mentioned description item question item analysis unit, a selection format and / or a text description format and / or a numerical description format equivalent to or related to the question item Related to the detailed questionnaire item generation unit that generates one or more question items, the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit, and the questionnaire item from which the questionnaire item is generated. A detailed questionnaire item recording section that records information in an appropriate format. By providing a questionnaire item generation device that can execute the questionnaire item generation unit and the questionnaire item recording unit only once, and can repeat the other processes once or more, the question items that can be expressed only in the text description format can be displayed. Support the process of enabling expression in a selection format.

Furthermore, by including the data regarding the profile of the questionnaire respondent in the question item generated by the detailed questionnaire item generating unit, the question respondent profile data and the description format question item analyzing unit are generated. By supporting appropriate combination of the analyzed results, we support the process of expressing question items that can be expressed only in the text description format in the selection format.

Furthermore, by further including a selection-form question item analysis unit for performing a statistical analysis process on the result of the question item in which the question-form is a selection form, the detailed-form questionnaire item generation unit includes the selection-form question. To make it possible to appropriately combine the analysis information of the question items in the selection format generated by the item analysis section, the profile data of the questionnaire respondents, and the analysis results generated by the description format question item analysis section. Supports the process of making question items that can only be expressed in text description format expressible in selection format.

Further, the description format question item analysis unit uses the morphological analysis method as the natural language processing and the clustering method as the statistical analysis processing to generate the structural information of the response result in the description format. Then
By appropriately using the information in the detailed questionnaire item generation unit, the process of making the question items that can be expressed only in the text description format expressible in the selection format is supported.

Furthermore, by repeating the process until all the questionnaire items that are finally generated are configured with the question items of the selection format except the question items regarding the profile of the answerer, only the question items of the selection format are included. Since the questionnaire form can be generated, a questionnaire item generation device that can efficiently (automatically) perform the analysis process after the questionnaire survey is realized.

In the present invention, one or more question items in the text description format and zero or more question items in the numeric description format and 0 are provided.
A questionnaire item generation unit that generates a questionnaire item composed of question items in the above selection format, a questionnaire item recording unit that records the content of the questionnaire item generated by the questionnaire item generation unit and information about the format in an appropriate format, In the questionnaire respondent identifier assigning unit that assigns an identifier to each questionnaire respondent, the questionnaire items recorded in the questionnaire item recording unit, the questionnaire items recorded in the detailed questionnaire item recording unit described below, and the questionnaire response result recording unit described later. Based on the information on the recorded questionnaire response results, a questionnaire enforcement section that generates a questionnaire form unique to each questionnaire respondent and enforces the questionnaire, and a questionnaire response result aggregation section that aggregates the response results in the questionnaire enforcement section , Aggregated by the questionnaire response result aggregation section For each questionnaire item recorded in the questionnaire response result recording section that records the aggregated result with the identifier of the questionnaire respondent in an appropriate format and the question item in the text description format, to the questionnaire result recorded in the corresponding questionnaire response result recording section, Natural language processing and /
Alternatively, by appropriately using the description format question item analysis unit that performs statistical analysis processing and the analysis information of each question item generated by the description format question item analysis unit, a selection format equivalent to or related to the question item And / or a detailed questionnaire item generator that generates one or more question items in text description format and / or numerical description format,
A detailed questionnaire item recording unit that records the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit and information about the questionnaire item from which the questionnaire item is generated in an appropriate format, By providing a questionnaire execution device that has a questionnaire item generation unit and a questionnaire item recording unit that are executed only once, and the other units are repeatedly executed once or more,
Information for identifying individual questionnaire respondents, answer information to question items, and information about question items that are the source of generating arbitrary question items are appropriately managed, and in the answer process of the questionnaire that is repeated, By providing a mechanism for generating a questionnaire form unique to a respondent at the time of answering, a questionnaire execution system that allows the questionnaire respondent to participate in the questionnaire survey at any time is provided.

Further, by appropriately managing the information on the question item that is the source of generating the arbitrary question item, it becomes the source of generating the arbitrary question item in the reply process of the questionnaire that is repeatedly performed. We provide a mechanism to generate a questionnaire form that does not include question items, and provide a questionnaire execution system that can reduce the burden on respondents.

Furthermore, by appropriately managing the information for identifying the individual respondents to the questionnaire and the answer information to the question items, the respondents have not yet answered at any point in the reply process of the repeated questionnaire. Provide a mechanism to generate a questionnaire form consisting of question items
Provide a questionnaire execution system that can reduce the burden on respondents.

Furthermore, by appropriately managing the information for identifying the individual respondents to the questionnaire and the answer information to the question items, the respondents have already answered at any point in the reply process of the repeated questionnaire. We provide a mechanism to input the question items in the questionnaire form in a non-editable format, and provide a questionnaire execution system that supports the respondent's answer process.

Further, in the description format question item analysis unit, by using the morphological analysis method as the natural language processing and the clustering method as the statistical analysis processing, the question items in the text description format can be efficiently selected. To provide a questionnaire execution system that can efficiently generate new question items (questionnaire form) by providing a mechanism to support the process of converting into question items.

Further, a questionnaire item including one or more text description type question items, 0 or more numerical description type question items, and 0 or more selection type question items is generated to generate a questionnaire item, and the generation is performed. The questionnaire items recorded in an appropriate format, the questionnaire items recorded above, the questionnaire items recorded in the detailed questionnaire item record described below,
And generate a questionnaire form based on the information of the questionnaire response result recorded in the questionnaire response result record described below, try the questionnaire, aggregate the response results in the trial of the questionnaire, and use the aggregated results in the appropriate format For each of the question items that are recorded and the question items are in the text description format, the natural language processing and / or the statistical analysis processing is applied to the response result of the corresponding questionnaire result aggregation, and the question items in the description format are analyzed. By appropriately using the analysis information of each question item generated by the analysis of the above-mentioned description item question item, a question in a selection form and / or a text description form and / or a numerical description form that are equivalent or related to the question item Generate one or more items to generate a detailed questionnaire item, and The content, the format, and the information about the questionnaire item from which the questionnaire item is generated are recorded in an appropriate format, the generation of the detailed questionnaire item and the recording of the detailed questionnaire item are executed only once, and other processing is performed. Provides a questionnaire item generation method in which a questionnaire item is generated by repeatedly executing it once or more.

Further, the questionnaire item generating section for generating a questionnaire item including one or more question items in the text description format and zero or more question items in the selection format and / or numerical description format, and the questionnaire item generation section. Questionnaire item recording section that records the information about the generated questionnaire item and its format in an appropriate format, questionnaire respondent identifier assigning section that assigns an identifier to each questionnaire respondent, and questionnaire recorded by the questionnaire item recording section Generate a questionnaire form unique to each questionnaire respondent based on the items, questionnaire items recorded in the detailed questionnaire item recording section described below, and information on the questionnaire response results recorded in the questionnaire response result recording section described below, Questionnaire enforcement department that conducts the questionnaire and the questionnaire In the enforcement department, a questionnaire enforcement counter that records the number of times the questionnaire is repeated for each questionnaire respondent identifier, and a questionnaire cooperation degree calculation unit that determines the questionnaire response cooperation degree for each questionnaire respondent identifier based on the output of the questionnaire enforcement counter And a respondent acquisition point recording unit that records the output of the questionnaire cooperation degree calculation unit for each questionnaire respondent, a questionnaire response result calculation unit that calculates the response results of the questionnaire enforcement unit, and the questionnaire response result calculation unit In the questionnaire response result recording section for recording the totalized results together with the identifiers of the questionnaire respondents in an appropriate format, and for each question item whose question item is in text description format, it is recorded in the corresponding questionnaire response result recording section. Responding to the questionnaire results in natural language A description format question items analyzer for performing physical and / or statistical analysis,
By appropriately using the analysis information related to each question item generated by the description format question item analysis unit, one or more selection formats equivalent to or related to the question item and / or
Alternatively, the detailed questionnaire item generation unit for generating question items in the text description format and / or numerical entry format, and the content, format, and the relevant questionnaire item of the detailed questionnaire item generated by the detailed questionnaire item generation unit It has a detailed questionnaire item recording section for recording information about a questionnaire item which is a generation source in an appropriate format, and the questionnaire item generating section and the questionnaire item recording section execute only once, and otherwise repeat once or more. (EN) Provided is a questionnaire execution device which improves motivation for a questionnaire by executing the questionnaire and obtains reliability of answer data.

Further, the questionnaire cooperation degree calculation unit makes the points for the second and subsequent answers at least equal to or higher than the points for the first and subsequent answers to motivate the second and subsequent questionnaires. Provided is a questionnaire execution device adapted to improve

Furthermore, the questionnaire execution device further enhances the motivation for answering the questionnaire by increasing the points given to each answer at least in proportion to the number of repetitions by the questionnaire cooperation degree calculating unit. I will provide a.

Furthermore, the respondent acquisition point recording unit manages points other than the points for the questionnaire. By converting and recording the output of the questionnaire cooperation degree calculation unit to the points, the questionnaire response acquisition points can be obtained. Provided is a questionnaire execution device capable of more effectively using the questionnaire.

Further, the questionnaire enforcement section further enhances the motivation for answering the questionnaire by providing information similar to the points calculated by the questionnaire cooperation degree calculating section (for example, the number of points for the answer) at the time of executing the questionnaire. Provide a questionnaire execution device.

Furthermore, a questionnaire item including one or more question items in the text description format and zero or more question items in the selection format and / or numerical description format is generated, and the content and the format of the generated questionnaire item are related. Information is recorded in an appropriate format, a respondent identifier is assigned to each questionnaire respondent, and the questionnaire items recorded above, the questionnaire items recorded in the detailed questionnaire item recording section described later, and the questionnaire response result recording section described later are recorded. Based on the information of the questionnaire response result recorded by
Conduct a questionnaire, count and record the number of times the questionnaire is repeated for each questionnaire respondent identifier in the execution of the questionnaire, and determine the questionnaire response cooperation level for each questionnaire respondent identifier based on the number of times the questionnaire is repeated The cooperation degree calculation section,
The degree of cooperation of the questionnaire is recorded for each respondent of the questionnaire to obtain points for respondents, and the result of the questionnaire response in the questionnaire enforcement is totaled, and the total result of the questionnaire response is recorded in an appropriate format together with the identifier of the questionnaire respondent. Then, for each question item in which the question item is a text description format, the corresponding questionnaire response result is subjected to natural language processing and / or statistical analysis processing to analyze the description format question item, and the description format question item analysis is performed. By appropriately using the analysis information related to each question item generated in, the question item in one or more selection format and / or text description format and / or numerical entry format that is equivalent to or related to the question item Generate and generate a detailed questionnaire item, and select from among the generated detailed questionnaire items. , Format, and information about the questionnaire item from which the questionnaire item is generated are recorded in an appropriate format. The generation of the questionnaire item and the recording of the questionnaire item are executed only once, and other times are repeated once or more. By doing so, it is possible to provide a questionnaire execution device that improves the motivation for the questionnaire and obtains the reliability of the response data.

[0040]

1 and 2 are block diagrams for explaining an embodiment of a questionnaire item generating device according to the present invention. In a questionnaire item generating unit 11, a profile survey item and target of a questionnaire target person are shown. Questionnaire items are generated so that question items for assignments are in a selection format, a numerical description format, or a text description format. There are various methods for creating questionnaire items, such as a method of creating question items by dialogue using an expert system and a method of constructing question items on a GUI using a method of visual programming. Although conceivable, the present invention does not limit the specific generation method,
Any method can be used as long as the questionnaire items can be appropriately generated.

FIG. 3 shows an example of the composition of questionnaire items.
Fig. 3 shows the questionnaire survey items regarding the new functions of the mobile phone, and the profile survey items (ID
= Q1, Q2), one numerical description format profile survey item (ID = Q3), two selection format question items (I
D = Q4, Q5) and one text description form question item (ID = Q6).

The questionnaire item recording unit 12 records the questionnaire items generated by the questionnaire item generating unit 11 in an appropriate format.

In the questionnaire trial unit 13, the questionnaire items recorded in the questionnaire item recording unit 12, the questionnaire items recorded in the detailed questionnaire item recording unit described later, and the questionnaire response results recorded in the questionnaire response result recording unit described later. Generate a questionnaire form based on the information of and try the questionnaire survey. Although various methods such as using the Internet can be used as the trial method of the questionnaire survey, the specific trial method is not limited in the present invention, and any method capable of appropriately conducting the questionnaire survey may be used.

The questionnaire response result totaling unit 14 totals the response results from the questionnaire trial unit 13. FIG. 4 shows an example of the result of totaling the results of the trial of the questionnaire using the questionnaire items shown in FIG. FIG. 4 is a compilation of the response results of 10 people, and in Q1 “male” = 1,
“Woman” = 2, in Q2 “student” = 1, “company employee” = 2,
“Unemployed” = 3, “Other” = 4, and in Q4 and Q5, “Yes” = 1 and “No” = 2. Also, in Q3, the age of the respondent is a numerical value, and in Q6, the answers to the question items are described in free text.

The questionnaire response result recording unit 15 records the totalized results aggregated by the questionnaire response result totaling unit 14 in an appropriate format.

Here, in the example shown in FIG. 2, in the selection format question item analysis unit 21, among the answer results recorded in the questionnaire response result recording unit 15, the results for the question items in the selection format are statistically compared. Apply processing. The answer result of the question item in the selection format can be expressed in the form of categorical data. FIG. 5 shows the result example shown in FIG. 4 expressed in the form of categorical data. The statistical processing method of categorical data is described in detail in, for example, “Quantification Theory and Data Processing” (Tsutomu Komazawa, Asakura Shoten) and “Introduction to Multivariate Analysis II” (Kazuto Kawaguchi, Morikita Publishing). There is. Here, as a simple example, FIG. 6 shows a cross table of Q1 and Q2 and Q4 and Q5 obtained by using the data shown in FIG.

Returning to FIG. 1 and FIG. 2, in the description format question item analysis unit 16, for each question item in which the question item is a text description format, the natural language is compared with the result of the corresponding questionnaire response result totaling unit 14. Perform processing and / or statistical analysis processing. Although various natural language processing methods and statistical analysis methods can be used, as an example, by applying Japanese morphological analysis to the free description sentence data of Q6 in FIG. 4, the part of speech is a noun. We show the process of extracting words, expressing each data as a vector based on the information of the noun words that compose each data, and applying the hierarchical clustering method to the generated vector.

FIG. 7 shows the results obtained by applying the Japanese morphological analysis to the free description sentence data of Q6 in FIG. 4 and extracting words whose part of speech is a noun from each sentence data. Next, a unique word existing in the whole sentence data is extracted, and if each word appears in each sentence data, 1 is set, and if not, 0 is obtained, and the appearance vector of the word in each document is obtained. It shows in FIG. Various statistical methods can be used by converting the sentence data into vector data as shown in FIG. 7. In this example, by applying the hierarchical clustering method to the sentence vector data shown in FIG. Is divided into several groups. As the hierarchical clustering method, the Ward method in which the similarity measure is the square distance is used (for details, the above-mentioned "Introduction to Multivariate Analysis II"). Figure 8
FIG. 9 shows a dendrogram as a result of applying the Ward method to the sentence vector data shown in FIG. From this dendrogram, sentence groups in an arbitrary hierarchy can be easily extracted. For example, the sentence group when the number of divisions is 2 is {A1, A4, A6, A7, A8, A10, A
5, A3}, {A2, A9} and the number of divisions is 5, the sentence groups are {A1, A4}, {A6, A7, A
8, A10}, {A5}, {A3}, {A2, A9}.

Originally, it seems natural to set the number of divisions to 3, but the result of the number of divisions 3 is different from what humans expected. However, this is a mathematically appropriate solution, and it is considered that the unnatural behavior of A3 and A5 is due to many constituent words.

The detailed questionnaire item generator 17 appropriately uses the information about each question item generated by the description format question item analyzer 16 and / or the selection format question item analyzer 21, One or more question items in the selection format and / or numerical description format and / or text description format that are equivalent or related to the question item are generated.

Similar to the questionnaire item generator 11,
As a method of generating questionnaire items, an GU is used by using an expert system and a method of generating question items by dialogue or a method of visual programming.
Although various methods such as a method of forming a question item on I can be considered, the present invention is not limited to a specific generation method, and any method can be used as long as a questionnaire item can be appropriately generated. However, this process is inherently difficult to perform automatically, and it is premised that the designer of the questionnaire item interactively generates the question item based on various information. Here, a process of inheriting the above example and generating a new question item from the result of the question item Q6 of FIG. 4 based on the information shown in FIGS.

First, FIG. 10 shows a categorical data format of sentence groups generated from the dendrogram of FIG. 9 when the number of divisions is 2 and 5, respectively.
First of all, a simple interpretation of the sentence group generated when the number of divisions is 2 shows that group 1 is "a group that wants to use the mobile phone to obtain class information and make various reservations. Group 2 can be thought of as "a group wanting to use a mobile phone as a substitute for a remote controller for home electric appliances". Based on this result
When each cell in the cross table of Q1-Q5 shown in (A) is further combined with the distribution situation when the number of divisions is 2, a result as shown in FIG. 11 is obtained. It is possible to obtain a hypothesis that "a person who does not use functions such as e-mails wants the mobile phone to have an additional function as a remote control for home electric appliances and the tendency is strong for men."

However, the problem here is Group 2
Although the respondents who belong to the group did not express the opinion that "the function of the remote control", it cannot be said that these data alone deny it. Therefore, in this case, a new question item in the selection format is created regarding the necessity of the "function of the mobile phone + remote control". On the other hand, the respondents who belong to Group 1 are considered to be "a group that wants to obtain class information and make various reservations using their mobile phones," as described above.
Judging from the above, it can be read that the sentence group obtained when the number of divisions is 3 or more has an unnatural result such that A3 and A5 form a single group.

Therefore, for the "function of reserving something", a question item in a selection format for asking the validity of the function and a question item in a text description format for asking a specific item are generated. In addition, since the “acquisition of class cancellation information” group {A1, A4} is very strong, it is possible to select whether or not “acquisition of class cancellation information” is effective for this group. Create a question item. As a result, the new question item shown in FIG. 12 is created.

In the detailed questionnaire item recording unit 18, the contents and format of the detailed questionnaire item generated by the detailed questionnaire item generating unit 17 and the information regarding the questionnaire item from which the questionnaire item is generated are appropriate. Record in any format. Further, based on the questionnaire items recorded in the detailed questionnaire item recording unit 18, the questionnaire trial unit 13, the questionnaire response result totaling unit 14, the questionnaire response result recording unit 15, and the selection format question item analyzing unit 21. The process of the description format question item analysis unit 16 is re-executed repeatedly, and this process is re-executed using the questionnaire items generated by the detailed questionnaire item generation unit 17 in the selection format. By repeating the processing until the processing is performed, the result of the desired question item can be obtained in the form of effective categorical data.

FIG. 13 is a block diagram for explaining one embodiment of the questionnaire execution device according to the present invention. In the questionnaire item generation unit 11, the profile survey items of the questionnaire target person are in the selection format and the question items for the target task are , Generate questionnaire items in the selection format and the description format. FIG. 14 shows an example of the generated question item.

The questionnaire item recording unit 12 records the questionnaire items generated by the questionnaire item generating unit 11 in an appropriate format. An example of recording the example of the question item shown in FIG. 14 in a table format is shown in FIGS. 15 and 16. In FIG. 15, an identification number (ID) uniquely assigned to each question item,
Question item type (type = 1: profile (selection format), 2: profile (description format), 3: question item (selection format), 4: question item (description format)), content of question item (contents), Reference identification number (refID, to the option table when the question item has options)
If there is no option, 0) is recorded, and the reference identification number (root ID, if there is no 0) to the identification number table of the question item from which the target question item is generated is recorded. In this example, it is premised that the numerical description format question item results are appropriately converted to categorical data,
No distinction is made between numerical description format and text description format, and they are treated as the same description format. Also, in FIG. 16, an option table for each question item is recorded. The reference identification number is recorded in the first line, and the identification numbers and contents of the option items are recorded in the second and subsequent lines.

In the questionnaire respondent identifier assigning section 31,
An identifier is assigned to each survey respondent.

In the questionnaire enforcement unit 32, the questionnaire items recorded in the questionnaire item recording unit 12 and the questionnaire items recorded in the detailed questionnaire item recording unit described later,
A questionnaire form specific to each questionnaire respondent is generated based on the information on the questionnaire reply result recorded in the questionnaire reply result recording unit, and the questionnaire is executed.

The questionnaire response result totaling unit 14 totals the response results from the questionnaire enforcement unit 32. 14
FIG. 17 shows an example of the result of totaling the results of once conducting the questionnaire using the questionnaire items shown in FIG. Figure 17
It is the answer result to the question items with the identification numbers 1 to 6 of the 10 respondents with the identification numbers 1 to 10. The answers to the question items with identification numbers 1, 2, 4, and 5 are indicated by the identification numbers of the corresponding options, and the answers to the question items with identification numbers 3 and 6 are described in text.

The questionnaire response result recording unit 15 records the totalized results aggregated by the questionnaire response result totaling unit 14 together with the identifiers of the questionnaire respondents in an appropriate format.

In the description format question item analysis unit 16, for each question item in which the question item is in the description format, the natural language processing and / or the statistics are performed on the questionnaire result recorded in the corresponding questionnaire response result recording unit 15. Perform analysis processing.

The detailed questionnaire item generator 17 appropriately uses information about each question item generated by the description-type question item analyzer 16 to select one or more items equivalent to or related to the question item. Generate a question item in the form or description.

The detailed questionnaire item recording unit 18 appropriately stores the content and format of the detailed questionnaire item generated by the detailed questionnaire item generating unit 17 and the information about the questionnaire item from which the detailed questionnaire item is generated. Record in any format. For example, as a result of the processing of the description format question item analysis unit and the detailed questionnaire item generation unit for the question item whose identification number is 6 in the questionnaire survey result shown in FIG. 17, the question item shown in FIG. 18 is newly generated. In this case, the contents recorded in the detailed questionnaire item section are as shown in FIG. The question item in FIG. 19 has the reference identification number 1 to the same identification number table, and all the above question items are generated from the question item with the identification number 6, so the identification number table has a table format as shown in FIG. Recorded in.

Next, a case will be described in which a new respondent is added at the time of conducting a questionnaire survey for the new question item shown in FIG. Since the answer to the question item shown in FIG. 14 has not been obtained from the new respondent, unless the answer result to the question item shown in FIG. 14 and the question item shown in FIG. Can not. However, it is not necessary to obtain the answer result for the question item that is the source of the arbitrary question item generation.

Therefore, the questionnaire form is shown in FIG.
The question item recorded in the identification number table of the question item indicated by 0 is excluded and generated. That is, in the questionnaire form for new respondents, the question item with the identification number 6 is excluded, and the identification numbers 1, 2, 3, 4, 5, 7, 8, 9,
It is composed of 10 question items. This can reduce the load on new respondents, for example, in the case of adding new respondents in the process of repeating the questionnaire survey (especially in the case where they are added after being repeated a plurality of times).

Next, a questionnaire survey for the new question items shown in FIG. 18 was carried out, and as a result, FIG.
A new question item as shown in FIG. 18 is generated, and when a questionnaire survey is further conducted on this question item, FIG.
The case where the questionnaire respondent who has not answered the questionnaire survey for the question item shown in (6) answers is explained.

The new question item shown in FIG. 21 is assumed to have been generated from question items having identification numbers 8 and 9. Further, since this question item is a question item of a selection format in which a plurality of items can be selected, a new identification number 5 is assigned, and the contents as shown in FIG. 22 are recorded. Survey respondents are shown in Figure 1.
If the user answers the question items shown in FIG. 4 but does not answer the question items shown in FIG. 18, unless the answer results for the question items shown in FIG. 21 and the question items shown in FIG. It cannot be used as data. However, it is not necessary to obtain the answer result for the question item that is the source of the arbitrary question item generation. Furthermore, it is not necessary to obtain an answer result for a question item for which an answer result has already been obtained by referring to the questionnaire answer result already recorded by the respondent.

Therefore, the questionnaire form is shown in FIG.
The question item in which the answer result of the respondent shown in FIG. 7 is recorded and the question item recorded in the identification number table of the question item shown in FIG. 23 are excluded and generated. That is, in this case, the questionnaire form has the identification number 1,
The question items 2, 3, 4, 5, 6, 8, and 9 are excluded, and the question items having identification numbers 7, 10, and 11 are included. As a result, even if the respondent did not answer in some repeating process in the process of repeating the questionnaire survey,
It is possible to reduce the load when respondents participate in the questionnaire survey again.

Furthermore, the excluded question items 1, 2,
When a question item that is not referred to as a source for generating an arbitrary question item is extracted from among 3, 4, 5, 6, 7, and 8, the identification numbers are 1, 2, 3 from FIG. 20 and FIG. ，
Question items 4 and 5 are extracted. These question items are
Since there is no redundancy with the new question item, the question form and the answer result are also generated to generate the questionnaire form. That is, in this case, the questionnaire form is composed of question items with identification numbers 1, 2, 3, 4, 5 and the answer result of the respondent, and question items with identification numbers 7, 10, 11. This allows the questionnaire respondents to recall the context of the entire questionnaire when the respondent participates in the questionnaire survey again, even if the respondent did not answer in some repeating steps during the repeated questionnaire survey. Can help.

FIG. 24 is a main part configuration diagram for explaining another example of the questionnaire execution device and the questionnaire execution method according to the present invention. In the questionnaire item generation unit 11, question item names and selection item names required for the questionnaire are shown. Enter to generate a questionnaire item. The generated questionnaire items are recorded in the questionnaire item recording unit 12 in an appropriate format. Figure 25
FIGS. 26 and 27 show examples of the questions shown in FIG.

In FIG. 26, the identification number (ID) uniquely assigned to each question item and the type of answer format (type = 1:
Selection format, type = 2: Description format (numerical value), type
= 3: description format (text)), question item (conte
nts), a reference item identification number (refID, 0 if there is no option) to the selection item table when the question item has options, and an identification number table of the question item from which the target question item is generated. Reference number to (rootI
D, 0) is recorded when there is no D. In addition, FIG.
Indicates a choice table for each question item. The reference identification number is recorded in the first line, and the identification numbers and contents of the selection items are recorded in the second and subsequent lines.

In the questionnaire respondent identifier assigning section 31,
Assign an identifier to each survey respondent.
In the present invention, the new respondent is made to register the profile of the respondent's name, age, contact information such as e-mail address, occupation, etc., and is given the respondent identifier. The questionnaire enforcement unit 32 also includes the questionnaire items recorded in the questionnaire item recording unit 12, the questionnaire items recorded in the detailed questionnaire item recording unit 18 described later, and the questionnaire response result information recorded in the questionnaire response result recording unit 15. In addition, a questionnaire form unique to each questionnaire respondent is generated and the questionnaire is executed. The questionnaire response result totaling unit 14 totals the response results from the questionnaire trial unit 32. FIG. 28 is a diagram showing an example of the result of totaling the results of once conducting a questionnaire using the questionnaire items shown in FIG. 25. The identification numbers for the 10 respondents with identification numbers 1 to 10 are 1 to 10. These are the answers to the question items up to 6. The answers to the question items with identification numbers 1, 2, 4, and 5 are indicated by the identification numbers of the corresponding options,
The answers to the question items with identification numbers 3 and 6 are described in text.

The questionnaire response result recording unit 15 records the totalized results totaled by the questionnaire response result totaling unit 14 together with the identifiers of the questionnaire respondents in an appropriate format. The description format question item analysis unit 16 performs natural language processing and / or statistical analysis processing on the questionnaire result recorded in the corresponding questionnaire response result recording unit 15 for each question item whose description item is description format. Give. The detailing questionnaire item generation unit 17 appropriately uses information about each question item generated by the description format question item analysis unit 16 so that one or more selection formats equivalent to or related to the question item or Generate question items in the description format.

In the detailed questionnaire item recording unit 18, the contents and format of the detailed questionnaire item generated by the detailed questionnaire item generating unit 17 and the information on the questionnaire item from which the questionnaire item is generated are appropriate. Record in any format. For example, as a result of processing of the description format question item analysis unit 16 and the detailed questionnaire item generation unit 17 for the question item with the identification number 6 of the questionnaire survey result shown in FIG. 28, the question item shown in FIG. When the detailed questionnaire item recording unit 18 is generated,
The contents recorded at are as shown in FIG. Since the question item in FIG. 30 has the reference identification number 1 to the same identification number table and all the above question items are generated from the question item with the identification number 6, the identification number table is as shown in FIG. It is recorded in a table format consisting of reference identification numbers of question items and question item identification numbers.

Next, a case where a new respondent is added at the time of conducting the questionnaire survey for the new question item shown in FIG. 29 will be exemplified. Since the answer to the question item shown in FIG. 25 has not been obtained from the new respondent, if the answer result to the question item shown in FIG. 25 and the question item shown in FIG. Can not. However, it is not necessary to obtain the answer result for the question item that is the source of the arbitrary question item generation. Therefore, the questionnaire form is generated by excluding the question items recorded in the question item identification number table shown in FIG. That is, the questionnaire form for new respondents is configured by the question items with the identification numbers 1, 2, 3, 4, 5, 7, 8, 9, 10 excluding the question item with the identification number 6. Thus, when a new respondent is added to the process of repeating the questionnaire survey (especially, when the new respondent is added after being repeated a plurality of times), the load on the new respondent can be reduced.

Furthermore, a questionnaire survey was conducted on the new question items shown in FIG. 29. As a result, FIG.
When a new question item as shown in Fig. 2 is generated and a questionnaire survey is further conducted on this question item,
The case where the questionnaire respondent who has not answered the questionnaire survey for the question items shown in 9 answers is illustrated.

The new question item shown in FIG. 32 is assumed to have been generated from question items having identification numbers 8 and 9. Further, since this question item is a question item of a selection format in which a plurality of items can be selected, a new identification number 5 is assigned, and the contents as shown in FIG. 33 are recorded. Figure 2 shows the respondents
If the user answers the question items shown in FIG. 5 but does not answer the question items shown in FIG. 29, unless the answer results for the question items shown in FIG. 32 and the question items shown in FIG. 29 are obtained. It cannot be used as analysis target data. However, it is not necessary to obtain the answer result for the question item that is the source of the arbitrary question item generation. Furthermore, it is not necessary to obtain an answer result for a question item for which an answer result has already been obtained by referring to the questionnaire answer result already recorded by the respondent.

Therefore, the questionnaire form is as shown in FIG.
The question item in which the answer result of the respondent shown in FIG. 8 is recorded and the question item recorded in the identification number table of the question item shown in FIG. 35 are excluded and generated. That is, in this case, the questionnaire form has the identification number 1,
The question items 2, 3, 4, 5, 6, 8, and 9 are excluded, and the question items having identification numbers 7, 10, and 11 are included. As a result, even if the respondent did not answer in some repeating process in the process of repeating the questionnaire survey,
It is possible to reduce the load when respondents participate in the questionnaire survey again.

Next, the questionnaire enforcement counter and the questionnaire cooperation degree calculation unit will be described. In FIG. 24, the questionnaire enforcement counter 33 counts the number of questionnaire enforcements for each questionnaire respondent identifier.
When the respondents are accepted by the questionnaire enforcement unit 32 and the respondent identifiers of the respondents are fixed, the counter and the questionnaire cooperation degree are initialized as shown in FIG.

The questionnaire enforcement section 32 repeats the questionnaire, and the questionnaire enforcement counter 33 is incremented by one each time an answer with the same respondent identifier is accepted. In FIG. 37, the respondent ID is 1,
2,3,4,6,9,10 respondents twice, respondent ID
Indicates that the respondents of 3, 5 responded three times and the respondents of other (ID7, 8) responded once.

The questionnaire cooperation degree calculating unit 34 calculates the questionnaire cooperation degree using a cooperation degree conversion table or a conversion formula corresponding to the frequency of the counter. FIG. 38 shows an example in which the degree of cooperation (points) given when the counter indicates at least the second time or later is set to be equal to or higher than that of the first time, and FIG. 39 is set so that points are increased each time it is repeated. It shows an example of the case. Here, the points (cooperation level) are rewards to the respondents, and may be considered as points to be paid according to the number such as credit card points or mileage when using an airplane, or to win a prize. You can think of it as a high probability.

In the above embodiment, the respondent acquisition point recording unit (management unit) 35, which records the output of the questionnaire cooperation degree calculation unit 34 by adding it to 1,000 points as acquisition points, is recorded.
have. The respondent acquisition point recording unit 35 is a table composed of respondent identifiers and points. Before answering the questionnaire, the counter is 0, so the points are also 0
Is. FIG. 40 shows an example of points of each respondent when using the conversion table of FIG. 39 when the questionnaire is repeated several times and the counter is in the state of FIG. 37.

The respondent whose respondent identifier (ID) is 5 (see FIG. 3)
ID = 5 in 7) replies 3 times, so if the conversion table in FIG. 39 is applied to this, 10 (first time)
+20 (second time) +30 (third time) = 60, which is 60 points (see ID = 5 in FIG. 40). In addition, although the respondent acquisition point recording unit 35 of the present embodiment describes only those related to the questionnaire that is currently being implemented, although this questionnaire system is actually used, it participates in this questionnaire. It may be implemented to record points, including those who did not.

The questionnaire enforcement section 32 has information corresponding to points calculated by the questionnaire cooperation degree calculation section 34 in order to increase the motivation of the respondents, for example, FIG. 38 or FIG.
Information such as is presented to the respondent. That is, the questionnaire enforcement unit 32 refers to the conversion table of FIG. 38 or FIG. 39, and the information (for example, for the respondent with ID = 5, the cooperation level is 20 (FIG. 38) or 30 (FIG. 39)). 36) is displayed at the time of enforcing the questionnaire.

The respondent-acquired point recording unit 35 manages the output of the questionnaire cooperation degree calculation unit 34, for example, for managing points other than questionnaire points such as credit card usage points and mileage points for using an aircraft. It can be reflected in the point data in the server (management unit other than the respondent acquisition point management unit) 37. The respondent acquisition point recording unit 35 is provided with a means for associating the output result of the questionnaire cooperation degree calculation unit 34 with a customer management table that manages credit card use points and mileage points and the point data thereof. There is.

First, the correspondence between the respondent identifier in the output of the questionnaire cooperation degree calculation unit 34 and the customer identifier managed in the customer management table is as follows: (1) The questionnaire respondent identifier is managed the same as the customer identifier. If
No means is required to match the two. (2) When the two are different, a correspondence table of respondent identifiers and customer identifiers is prepared to check the identity, and the questionnaire respondent identifiers are used as customer identifiers based on the correspondence table. It can be realized by converting to.

Next, the degree of cooperation for the questionnaire response is
Managing points other than questionnaire points (respondents) Acquisition points To reflect in the point data in the recording section, (1) Questionnaire cooperation level manages points other than questionnaire points (respondents) acquired points If it has the same unit as the point data in the recording section, there is no need for a means for matching them. (2) If they do not match, it can be realized by preparing a conversion table for both as shown in FIG.

[0089]

According to the present invention, a questionnaire item generating unit for generating a questionnaire item including one or more description type question items, 0 or more numerical description type question items, and 0 or more text selection type question items. A questionnaire item recording unit that records the questionnaire items generated by the questionnaire item generating unit in an appropriate format, a questionnaire item recorded by the questionnaire item recording unit, a questionnaire item recorded by the detailed questionnaire item recording unit described later, and A questionnaire trial part that creates a questionnaire form based on the information of the questionnaire response results recorded in the questionnaire response result recording part and a total of the response results in the questionnaire trial part and the questionnaire response result totaling part And, the result of totaling by the questionnaire response result totaling section is in an appropriate format. A questionnaire response result recording section to record, and a description format question item analysis section that performs natural language processing and / or statistical analysis processing on the results of the corresponding questionnaire result aggregation section for each question item in text description format, and Description format By properly using the analysis information of each question item generated by the question item analysis unit,
Selection format equivalent to or related to the question item and /
Alternatively, the detailed questionnaire item generation unit that generates one or more question items in the numerical description format and / or the text description format, and the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit and the questionnaire It includes a detailed questionnaire item recording unit that records information about the questionnaire item that is the source of item generation in an appropriate format, executes the questionnaire item generating unit and the questionnaire item recording unit only once, and performs other processing once. By providing the questionnaire item generation device that can be repeatedly executed, it is possible to support the process of making it possible to express the question item in the selection form, which can be expressed only in the text description form.

Further, by including the question items related to the profile of the questionnaire respondent in the question items generated by the questionnaire item generating unit, the detail data of the questionnaire respondent and the question item analyzing unit in the description format are included in the detailed questionnaire item generating unit. By making it possible to appropriately combine and use the analysis results generated in, it is possible to support the process of making it possible to express the question items that could be expressed only in the text description format in the selection format.

In addition, by further including a selection-format question-item analysis unit that performs statistical analysis processing on the results of the question items in which the question-items are in the selection format, in the detailed questionnaire item generation unit, the selection-format question-item analysis unit is included. The analysis information of the question items in the selection format generated in the above, the profile data of the questionnaire respondents, and the description format The text description format can be used by appropriately combining the analysis results generated by the question item analysis unit. It is possible to support the process of making it possible to express a question item that could be expressed only in a selective form.

In the description format question item analysis unit, by using the morphological analysis method as the natural language processing and the clustering method as the statistical analysis processing, structural information of the answer result in the description format is generated, and the information thereof is generated. Can be appropriately used in the detailed questionnaire item generation unit, thereby supporting the process of allowing the question item, which can be expressed only in the text description format, to be expressed in the selection format.

By repeating the process until all the questionnaire items finally generated are configured with the question items in the selection format except the question items related to the answerer's profile, the questionnaire form including only the question items in the selection format Therefore, it is possible to realize a questionnaire item generation device that can efficiently (automatically) perform the analysis process after the questionnaire survey.

Questionnaire answers that are repeatedly performed by appropriately managing information for identifying individual questionnaire respondents, answer information to question items, and information about question items that are the source of generating arbitrary question items In the process
By providing a mechanism for generating a questionnaire form unique to a respondent at an arbitrary point of time, it is possible to provide a questionnaire implementation system that enables a questionnaire respondent to participate in a questionnaire survey at an arbitrary point of time.

Furthermore, by appropriately managing the information on the question item that is the source of the arbitrary question item, the question that is the source of the arbitrary question item in the process of answering the questionnaire that is repeatedly performed. By providing a mechanism for generating a questionnaire form that does not include items, it is possible to provide a questionnaire implementation system that can reduce the burden on respondents.

Furthermore, by appropriately managing the information for identifying the individual respondents to the questionnaire and the answer information to the question items, the respondents have not yet answered at any point in the reply process of the repeated questionnaire. By providing a mechanism for generating a questionnaire form consisting of question items, it is possible to provide a questionnaire implementation system that can reduce the burden on respondents.

Furthermore, by appropriately managing the information for identifying the individual respondents to the questionnaire and the answer information to the question items, the respondents have already answered at any point in the reply process of the repeated questionnaire. By providing a mechanism in which the question items being asked are entered in a non-editable form in the questionnaire form, it is possible to provide a questionnaire implementation system that supports the answering process of the respondents.

Furthermore, in the description format question item analysis unit, by using the morphological analysis method as the natural language processing and the clustering method as the statistical analysis processing, the question items in the text description format can be efficiently selected. By providing a mechanism that supports the process of converting into question items, it is possible to provide a questionnaire implementation system that can efficiently generate new question items (questionnaire form).

As described above, according to the embodiment shown in FIG. 24, it is possible to obtain higher points than the first time for the second and subsequent questionnaire responses in which the motivation of the respondent is expected to decrease. , Increase motivation and the reliability of the resulting data.

Generally, companies have high needs for grasping trends of customers who manage them. For this reason, points are being rewarded to customers according to the frequency of use (or the amount of use) of credit cards and passenger planes, and efforts are being made to understand such points. What is obtained there is record data of purchases and uses, but since the present invention can additionally carry out a questionnaire, it becomes possible to grasp abundant customer trends.

The points that can be received at the time of reply are clearly indicated even if the points are different depending on the questionnaire, so that the motivation of the respondents increases, and the questionnaire implementer increases the points given to increase the response rate. You can

[Brief description of drawings]

FIG. 1 is a configuration diagram for explaining an embodiment of a questionnaire item generation device according to the present invention.

FIG. 2 is a configuration diagram for explaining another example of the questionnaire item generating device.

FIG. 3 is a diagram showing a configuration example of questionnaire items.

FIG. 4 is a diagram showing an example of aggregating reply results.

FIG. 5 is a diagram expressing the result example shown in FIG. 4 in the form of categorical data.

6 is a diagram showing a cross table of Q1 and Q2, and Q4 and Q5 using the data shown in FIG.

FIG. 7 is a diagram showing a result of applying a morphological analysis to extract a word whose part of speech is a noun from each sentence data.

FIG. 8 is a diagram in which an appearance vector of a word in each document is obtained.

9] War for the sentence vector data shown in FIG.
It is a figure which shows the dendrogram of the result of applying the d method.

FIG. 10 shows that the number of divisions is 2 and 5 from the dendrogram of FIG.
In the case of, the sentence groups respectively generated are represented in the form of categorical data.

FIG. 11 is a diagram showing a result of synthesizing distribution states when the number of divisions is two.

FIG. 12 is a diagram showing an example when a new question item is created.

FIG. 13 is a configuration diagram for explaining an example of the questionnaire execution device according to the present invention.

FIG. 14 is a diagram showing an example of generated question items.

FIG. 15 is a diagram showing an example in which the example of the question items shown in FIG. 14 is recorded in a table format.

FIG. 16 is a diagram showing an example in which the example of the question items shown in FIG. 14 is recorded in a table format.

FIG. 17 is a diagram showing an example of the result of totaling the results of once conducting a questionnaire using the questionnaire items shown in FIG.

FIG. 18 is a diagram showing a result of processing by the description format question item analysis unit and the detailed questionnaire item generation unit for the question item with the identification number 6 of the questionnaire survey result shown in FIG. 17;

FIG. 19 is a diagram showing the contents recorded in the detailed questionnaire item portion when the question item shown in FIG. 18 is newly generated.

FIG. 20 is a diagram showing an identification number table.

FIG. 21 is a diagram showing a new question item.

FIG. 22 is a diagram showing new question items.

FIG. 23 is a diagram showing question items recorded in a question item identification number table.

FIG. 24 is a main part configuration diagram for explaining an example of a questionnaire implementation device and a questionnaire implementation method according to the present invention.

FIG. 25 is a diagram showing an example of a question in a questionnaire.

FIG. 26 is a diagram in which a question example of the questionnaire shown in FIG. 25 is recorded in a table format.

FIG. 27 is a diagram in which an option table for the question items of the questionnaire shown in FIG. 25 is recorded.

FIG. 28 is a diagram showing an example in which the results of conducting a questionnaire once using the questionnaire items shown in FIG. 25 are totaled.

FIG. 29 is a diagram showing an example of a case where a question item is newly generated.

FIG. 30 is a diagram in which the questions shown in FIG. 29 are recorded in a table format.

FIG. 31 is a diagram showing a question item identification number table.

FIG. 32 is a diagram showing newly created items.

FIG. 33 is a diagram showing the question example of FIG. 32 in a tabular form.

34 is a diagram showing the new question shown in FIG. 32. FIG.

FIG. 35 is a diagram showing an identification number table for question items.

FIG. 36 is a diagram showing an initialized state of a counter and a questionnaire cooperation level.

FIG. 37 is a diagram showing the number of times a respondent answers.

FIG. 38 is a diagram showing an example of setting the degree of cooperation (points) of the respondent.

FIG. 39 is a diagram showing another example of setting the cooperation level (points) of the respondent.

FIG. 40 is a diagram showing an example of points of each respondent when the conversion table of FIG. 39 is used.

FIG. 41 is a diagram showing an example of a conversion table with a point system in another system.

[Explanation of symbols]

11 ... Questionnaire item generation unit, 12 ... Questionnaire item recording unit, 13 ... Questionnaire trial unit, 14 ... Questionnaire answer result counting unit, 15 ... Questionnaire answer result recording unit, 16 ...
Description format Question item analysis unit, 17 ... Detailed questionnaire item generation unit, 18 ... Detailed questionnaire item recording unit, 21 ... Selection format question item analysis unit, 31 ... Questionnaire respondent identifier assigning unit, 32 ... Questionnaire enforcement unit, 33 … Questionnaire enforcement counter, 34… Questionnaire cooperation degree calculation unit, 35…
Respondent acquisition point recording unit, 36 ... Display unit, 37 ... Point management server of other system.

Claims (13)

[Claims] 1. A questionnaire item generation unit for generating a questionnaire item including one or more text description type question items, 0 or more numerical description type question items, and 0 or more selection type question items, and the questionnaire item. Questionnaire item recording unit that records the questionnaire items generated by the generation unit in an appropriate format, questionnaire items recorded by the questionnaire item recording unit, questionnaire items recorded by the detailed questionnaire item recording unit described below, and questionnaire response result recording described below Generate a questionnaire form based on the information of the questionnaire response results recorded in the department, and try the questionnaire, the questionnaire trial part, the questionnaire response result totaling part that totals the response results in the questionnaire trial part, and the questionnaire response result Questionnaire response results that record the aggregation results aggregated by the aggregation section in an appropriate format A recording unit, a description format question item analysis unit that performs natural language processing and / or statistical analysis processing on the response results of the corresponding questionnaire result aggregation unit for each of the question items in which the question items are in text description format; Description format Question items in the selection format and / or text description format and / or numerical description format that are equivalent or related to the relevant question item by properly using the analysis information of each question item generated by the question item analysis unit. The detailed questionnaire item generation unit that generates one or more items, and the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit and the information about the questionnaire item that is the source of generating the questionnaire item. A detailed questionnaire item recording section for recording in an appropriate format, and generating the detailed questionnaire item A questionnaire item generating device characterized in that a questionnaire item is generated by executing a copy section and a detailed questionnaire item recording section only once, and repeating other processes once or more. 2. The detailed questionnaire item generation device according to claim 1, further comprising a selection form question item analysis unit that performs statistical analysis processing on a result of the question item in which the question item is a selection form, A questionnaire item generation device characterized in that analysis information of a question item in a selection format generated by a question item analysis unit is also used in an appropriate format. 3. A questionnaire item generation unit for generating a questionnaire item comprising one or more description type question items, 0 or more numerical description type question items, and 0 or more selection type question items, and the questionnaire item. In the questionnaire item recording unit that records the content and the format of the questionnaire item generated by the generation unit in an appropriate format, the questionnaire respondent identifier assigning unit that assigns an identifier to each questionnaire respondent, and the questionnaire item recording unit A questionnaire form specific to each respondent is created based on the recorded questionnaire items, the questionnaire items recorded in the detailed questionnaire item recording section below, and the questionnaire response result information recorded in the questionnaire response result recording section below. ,
Questionnaire enforcement department that implements the questionnaire, questionnaire response result aggregation unit that aggregates the response results in the questionnaire enforcement unit, and the aggregation results aggregated in the questionnaire response result aggregation unit together with the identifiers of the questionnaire respondents in an appropriate format Natural language processing and / or statistical analysis processing is performed on the questionnaire results recorded in the questionnaire response result recording section corresponding to the questionnaire response result recording section and the question items whose question items are in text description format. By appropriately using the description format question item analysis unit to perform and the analysis information about each question item generated by the description format question item analysis unit, the selection format and / or the text description equivalent to or related to the question item 1 question item in format and / or numerical description format
Appropriate information about the detailed questionnaire item generation unit that generates one or more, the content and format of the detailed questionnaire item generated by the detailed questionnaire item generation unit, and the information on the questionnaire item from which the relevant questionnaire item is generated. And a detailed questionnaire item recording section for recording in a format,
A questionnaire execution device characterized in that the questionnaire item generating unit and the questionnaire item recording unit are executed only once, and the other units are repeatedly executed once or more. 4. The questionnaire execution device according to claim 3, wherein the questionnaire form generated by the questionnaire enforcement unit generates an arbitrary questionnaire item recorded in the detailed questionnaire item recording unit. A questionnaire execution device characterized in that it does not include the questionnaire item that is the source of the questionnaire. 5. The questionnaire execution device according to claim 3 or 4, wherein the questionnaire form generated by the questionnaire enforcement unit for each respondent is a questionnaire item for which the respondent has already answered. A questionnaire execution device characterized in that the answer result recorded in the answer result recording unit is inputted in a non-editable format. 6. A questionnaire item comprising one or more text description type question items, 0 or more numerical description type question items, and 0 or more selection type question items is generated to generate a questionnaire item, and the questionnaire item is generated. Record the generated questionnaire items in an appropriate format, based on the questionnaire items recorded above, the questionnaire items recorded in the detailed questionnaire item record below, and the questionnaire response result information recorded in the questionnaire response result record described below. Produces
The questionnaire is tried, the results of the questionnaire trial are aggregated, the aggregated results are recorded in an appropriate format, and the corresponding questionnaire results are aggregated for each question item in which the question item is in text description format. Natural language processing and / or statistical analysis processing is applied to the answer result to analyze the question item in the description format, and the analysis information of each question item generated in the analysis of the question item in the description format is appropriately used. Then, one or more question items in a selection format and / or a text description format and / or a numerical description format that are equivalent to or related to the question item are generated to generate a detailed questionnaire item, and the generated detailed questionnaire item is generated. Record the content, format, and information about the questionnaire item from which the questionnaire item is generated in an appropriate format, A method for generating a questionnaire item, wherein the generation of the detailed questionnaire item and the recording of the detailed questionnaire item are executed only once, and the other process is repeatedly executed once or more to generate the questionnaire item. 7. A questionnaire item generation unit for generating a questionnaire item comprising one or more question items in a text description format and zero or more question items in a selection format and / or a numerical description format, and a questionnaire item generation section. Questionnaire item recording section that records information about the content and format of the questionnaire item in a suitable format, questionnaire respondent identifier assigning section that assigns an identifier to each questionnaire respondent, and questionnaire item recorded in the questionnaire item recording section Based on the questionnaire items recorded in the detailed questionnaire item recording section described below and the information on the questionnaire response results recorded in the questionnaire response result recording section described below, a questionnaire form unique to each questionnaire respondent is generated, A questionnaire enforcement department that enforces
In the questionnaire enforcement department, a questionnaire cooperation counter that records the number of times the questionnaire is repeated for each questionnaire respondent identifier, and a questionnaire cooperation degree calculation that determines the questionnaire reply cooperation degree for each questionnaire respondent identifier based on the output of the questionnaire execution counter Section, an answerer acquisition point recording section that records the output of the questionnaire cooperation degree calculation section for each questionnaire respondent, a questionnaire response result aggregation section that aggregates the response results in the questionnaire enforcement section, and the questionnaire response result aggregation The questionnaire response result recording unit that records the aggregation results aggregated by the department together with the questionnaire respondents' identifiers in an appropriate format, and the questionnaire response result recording unit that corresponds to each question item in which the question items are in text description format For the survey results Equivalent to the question item by appropriately using the description format question item analysis unit that performs natural language processing and / or statistical analysis processing and the analysis information about each question item generated by the description format question item analysis unit. Alternatively, a detailed questionnaire item generation unit that generates question items in one or more related selection formats and / or text description formats and / or numerical entry formats, and a detailed questionnaire that is generated by the detailed questionnaire item generation unit. And a detailed questionnaire item recording section for recording information on the content and format of the item and the questionnaire item from which the questionnaire item is generated, in an appropriate format, and the questionnaire item generating section and the questionnaire item recording section A questionnaire execution device characterized by being executed only once and other processes being executed once or more . 8. The method according to claim 7, wherein the questionnaire cooperation degree calculation unit makes the point for the answer at least the second and subsequent repetitions equal to or higher than the point for the first answer. Questionnaire execution device. 9. The point which the questionnaire cooperation degree calculation section gives to each reply at least in proportion to the number of times of repetition increases.
Questionnaire execution device described in. 10. The answerer acquisition point recording unit also manages points other than points for a questionnaire, and converts the output of the questionnaire cooperation degree calculation unit and records the points. Item 10. The questionnaire execution device according to any one of items 7 to 9. 11. The questionnaire enforcement section provides information similar to the points calculated by the questionnaire cooperation degree calculation section at the time of conducting the questionnaire.
11. The questionnaire execution device according to any one of 1 to 10. 12. A questionnaire item including one or more question items in a text description format and zero or more question items in a selection format and / or a numerical description format is generated, and information about the content and the format of the generated questionnaire item. Is recorded in an appropriate format, and a respondent identifier is assigned to each questionnaire respondent, and the questionnaire items recorded above, the questionnaire items recorded in the detailed questionnaire item recording section described later, and the questionnaire response result recording section described later Based on the information on the recorded questionnaire response results, a questionnaire form unique to each questionnaire answerer is generated, the questionnaire is enforced, and the number of times the questionnaire is repeated for each questionnaire respondent identifier is counted and recorded. , Based on the number of times the questionnaire is repeated, G. The questionnaire cooperation degree is determined for each respondent identifier and used as a questionnaire cooperation degree calculation unit. The questionnaire cooperation degree is recorded for each questionnaire respondent to obtain answerer acquisition points, and the questionnaire answer result in the questionnaire enforcement is displayed. Tally, record the questionnaire response results in a suitable format together with the identifiers of the questionnaire respondents, for each question item in which the question item is a text description format, for the corresponding questionnaire response results,
Equivalent to the question item by analyzing the description item question item by performing natural language processing and / or statistical analysis processing and appropriately using the analysis information about each question item generated by the description item question item analysis. Alternatively, a detailed questionnaire item is generated by generating question items in one or more related selection formats and / or text description formats and / or numerical entry formats, and the content, format, and Information about a questionnaire item that is a source of generating the questionnaire item is recorded in an appropriate format, the generation of the questionnaire item and the recording of the questionnaire item are executed only once, and the other operations are repeatedly executed once or more. How to execute the questionnaire. 13. A computer-readable recording medium in which a program for executing the questionnaire execution method according to claim 12 is recorded.